Hi!
Recently I started to design a bass tube preamp.
Firstly I though the Alembic F2-B as the way to go, but reading my tubes books and primarly the "Valve Wizard" stuffs I ended to draw a design with DC coupled cathode follower. Certainly i wish a clean sound. So, my doubt is if the circuit will distort heavely my clean bass signal... I don't want an overdriven sound... I know that earlies bassmans (fender) first stages are Dc Cathode followers...
I wish to know your opinions.
Do you think it will work??
Recently I started to design a bass tube preamp.
Firstly I though the Alembic F2-B as the way to go, but reading my tubes books and primarly the "Valve Wizard" stuffs I ended to draw a design with DC coupled cathode follower. Certainly i wish a clean sound. So, my doubt is if the circuit will distort heavely my clean bass signal... I don't want an overdriven sound... I know that earlies bassmans (fender) first stages are Dc Cathode followers...
I wish to know your opinions.
Do you think it will work??
Attachments
Your tone control stack is a genuine Fender one. I've never seen a Fender amp with a cathode follower in front of it, though. So, why not use the 1st 12AX7 half as the input section, as shown, immediately followed by the TCS, followed by the 2nd half as a VAS? Then you could choose a beefier valve as the output CF, which might be useful, especially if you intend to apply an output transformer.
Best regards!
Best regards!
(Es para proteger el triodo que hace de seguidor de cátodo al enceder el preamp...)
It is to protect de tube from arcing when the preamp is turning on..
Look this early fender bassman (the bass input): https://s-media-cache-ak0.pinimg.com/originals/5d/50/36/5d50362a47e4336d63745b28141dfd30.jpg
Ampeg did the same on the SVT with a 12DW7. I guess 12DW7 tube will be a better option for a DC coupled CF, but I have a bunch of 12AX7 on my desk..
Well,.. I'm not mentioned what I pretend with this design is to emphathis the effect of DC coupled CF, that is compress the upper portion of the signal while the bottom portion is untouched. This, in theory, produce second order harmonics.. Is for this that I put two DC couple CF...
Look at the 2º paragraphe that I wrote.
Rather use a 12AU7.I guess 12DW7 tube will be a better option for a DC coupled CF, but I have a bunch of 12AX7 on my desk..
Shure, but the resistances in your schematic aren't very apropriate for a AX.
Mona
Did you mean to the 68K resistors?..
I choose them beacouse I don't want tons of gain.. (A tipycal 12ax7 stage has aprox. a gain of 60 with a 100K load resistor)
And a bandaxall tone stack.
Have a look at the Ampeg B15 (a classic ultra clean preamp - the grits in the PI) or Fred Nachbaur's Dogzilla
> why the 10k+diode networks?
Merlin "Valve Wizard" advocates this to protect the grid. I do not think that is necessary, but probably makes zero difference, and it is only a few pennies.
> cathode follower
But this is NOT a cathode follower!
It is just-barely a CF when Fender's values are used. Several authors have pointed out that small signal swing drive the CF's grid positive of its cathode. Large grid current flows. The gain-stage is *heavily* loaded for part of the signal swing. It is a kinky thing.
Here you have used non-Leo values. He had 680 as the first cathode bias, and 100K in first plate and second cathode, which is the condition where it comes out of cathode-follower mode on signal swing. You have put the 1.5K found in most other Leo stage designs with a new lower 68K load on both stages.
A pencil run-through says the "CF" must idle like a 20K resistor. A 12AX7 will NOT sink to such a low resistance in the negative grid range, or even at any reasonable positive grid voltage.
(I have assumed a 300V supply, but ratios will not be much different 150V-450V.)
Data-check: 12AX7 plot shows 68Vp-k 3.4mA far out in the positive grid range.
So V1b works as a *diode*, not a cathode follower. Apparent waste of money and heat. It throws heavy load on V1a voltage-amp stage. V1a works as-if the plate resistor was 34K(!) to a 150V supply. Voltage-gain near 10. Max un-distorted output <25V peak.
IMHO, go steal a Black/SilverFace guitar amp and use larger caps. The 5F6a "BassMan" was not a popular BASS amp; it gained fame when guitarists adopted the unloved beast. (The Alembic "is" the front-end of a Black/SilverFace, and good for many purposes. Its main shortcoming is that it was not intended to drive long wires or low impedances.)
Merlin "Valve Wizard" advocates this to protect the grid. I do not think that is necessary, but probably makes zero difference, and it is only a few pennies.
> cathode follower
But this is NOT a cathode follower!
It is just-barely a CF when Fender's values are used. Several authors have pointed out that small signal swing drive the CF's grid positive of its cathode. Large grid current flows. The gain-stage is *heavily* loaded for part of the signal swing. It is a kinky thing.
Here you have used non-Leo values. He had 680 as the first cathode bias, and 100K in first plate and second cathode, which is the condition where it comes out of cathode-follower mode on signal swing. You have put the 1.5K found in most other Leo stage designs with a new lower 68K load on both stages.
A pencil run-through says the "CF" must idle like a 20K resistor. A 12AX7 will NOT sink to such a low resistance in the negative grid range, or even at any reasonable positive grid voltage.
(I have assumed a 300V supply, but ratios will not be much different 150V-450V.)
Data-check: 12AX7 plot shows 68Vp-k 3.4mA far out in the positive grid range.
So V1b works as a *diode*, not a cathode follower. Apparent waste of money and heat. It throws heavy load on V1a voltage-amp stage. V1a works as-if the plate resistor was 34K(!) to a 150V supply. Voltage-gain near 10. Max un-distorted output <25V peak.
IMHO, go steal a Black/SilverFace guitar amp and use larger caps. The 5F6a "BassMan" was not a popular BASS amp; it gained fame when guitarists adopted the unloved beast. (The Alembic "is" the front-end of a Black/SilverFace, and good for many purposes. Its main shortcoming is that it was not intended to drive long wires or low impedances.)
Attachments
Last edited:
I have been messing around with making a preamp. I used the typical AB763 front end and a unity gain LT1358 as an output buffer to drive heavy loads. The LT1358 has regulated +/- supply.
Simple and sounds great.
I also like the Ampeg B15 front end.
Simple and sounds great.
I also like the Ampeg B15 front end.
Attachments
I know if I design it, it won't sound as good to you as your own design.
But I must suggest a path I think will lead to better results.
> typical AB763 front end and a unity gain LT1358
Just what I was thinking.
> I have a bunch of 12AX7 on my desk..
So use that instead of the '1358 chip.
It *is* important to know where you are going. I will *assume* you may want to feed basic Power Amps. So 2V output. And some modern Power Amps are low-Z, and you may need more than one to fill a stadium. So say >2Vrms in 10K load.
The AB763 front-end is THE most popular in the world because it gives a light flavor without overloading. I've made the cathode cap large for bass. I made the cap switchable for a lower-gain more-clean option.
Some of the tone-stack values used for guitar-tone may want to be 1.5X-2X bigger for bass. You will have to experiment; NO way to "design" what the ear really hears.
It usually works best if the sensitivity at Volume pot output is 50mV-100mV. Lower is hissy and higher makes the stage before work too hard. No shock that this is about what you get in the AB763.
Now we need to boost 50mV to 2V (gain of 40) and deliver >0.28mA in 10K load. The final tube needs to idle 5V-10X that signal current, 1.5mA-3mA. This is hot for one 12AX7 but "I have a bunch of 12AX7", so put two sections in parallel. That has some power output but not much gain, and ~~20K output impedance.
So add a gain-stage in front. Now too much gain, so use it for NFB to make this section clean and low impedance. This mildly complicates the output cap, but a second cap can be low-voltage so large yet cheap/small.
Sorry, no Cathode Follower. CFs have uses but cascoded CC stages can often do as well or better.
Do Not Forget B+ Filtering! Four gain-stages on one B+ node want to motor-boat. (If some are CF, maybe not.) Stay with 2 stages per node. The final stage has huge signal and does not need extra filtering. The first two stages need "clean" (the vol-recovery near as much as the input stage) and can share a node.
I see no good reason for the UnBal jack to switch the BAL network out. You might want to use both at once. (One to on-stage power amp, one to house console.)
But I must suggest a path I think will lead to better results.
> typical AB763 front end and a unity gain LT1358
Just what I was thinking.
> I have a bunch of 12AX7 on my desk..
So use that instead of the '1358 chip.
It *is* important to know where you are going. I will *assume* you may want to feed basic Power Amps. So 2V output. And some modern Power Amps are low-Z, and you may need more than one to fill a stadium. So say >2Vrms in 10K load.
The AB763 front-end is THE most popular in the world because it gives a light flavor without overloading. I've made the cathode cap large for bass. I made the cap switchable for a lower-gain more-clean option.
Some of the tone-stack values used for guitar-tone may want to be 1.5X-2X bigger for bass. You will have to experiment; NO way to "design" what the ear really hears.
It usually works best if the sensitivity at Volume pot output is 50mV-100mV. Lower is hissy and higher makes the stage before work too hard. No shock that this is about what you get in the AB763.
Now we need to boost 50mV to 2V (gain of 40) and deliver >0.28mA in 10K load. The final tube needs to idle 5V-10X that signal current, 1.5mA-3mA. This is hot for one 12AX7 but "I have a bunch of 12AX7", so put two sections in parallel. That has some power output but not much gain, and ~~20K output impedance.
So add a gain-stage in front. Now too much gain, so use it for NFB to make this section clean and low impedance. This mildly complicates the output cap, but a second cap can be low-voltage so large yet cheap/small.
Sorry, no Cathode Follower. CFs have uses but cascoded CC stages can often do as well or better.
Do Not Forget B+ Filtering! Four gain-stages on one B+ node want to motor-boat. (If some are CF, maybe not.) Stay with 2 stages per node. The final stage has huge signal and does not need extra filtering. The first two stages need "clean" (the vol-recovery near as much as the input stage) and can share a node.
I see no good reason for the UnBal jack to switch the BAL network out. You might want to use both at once. (One to on-stage power amp, one to house console.)
Attachments
Last edited:
Good idea, since he has 12AX7's to burn, put two triodes in parallel.
Those two capacitors in series at the output, not so good.The anodes has to make big swings at low frequencies to get the output at the right amplitude.The feedback resistor can be connected there without a capacior in between, there is DC ground via the transformer winding.
Only little problem, the 4µ7 has to be high voltage now.
Mona
Those two capacitors in series at the output, not so good.The anodes has to make big swings at low frequencies to get the output at the right amplitude.The feedback resistor can be connected there without a capacior in between, there is DC ground via the transformer winding.
Only little problem, the 4µ7 has to be high voltage now.
Mona
Attachments
You can change the flavor of the ab763 circuit quite a bit by using different tubes. I was testing several brands of 12ax7s and I can get 1.6% THD with say a GE 12AX7 or 5% THD with Sylvania. I measure @ 2v RMS out. I FFT every tube and label it to try and quantify sonic attributes. RCA tubes are fairly consistent and have a nice balanced FFT spectrum, it is no wonder they are favored in this circuit. FWIW LTspice says 2.2%THD for the Ayumi model.
I tried a lot of designs and the AB763 always sounds great to me so that is what I stuck with.
All great advice from everyone here for the OP to think about.
I tried a lot of designs and the AB763 always sounds great to me so that is what I stuck with.
All great advice from everyone here for the OP to think about.
> two capacitors in series at the output, not so good.The anodes has to make big swings
We got 300V supply and only need a few Volts at the output jack. The proposed 0.47uFd is 10K at 35Hz. The both in series, 10K @ 40Hz. So the bottom note of bass needs ~~4V plate swing for 2V output. Not serious.
I started that way because the NFB loop leaks DC voltage to the output. This is rude. Hence the second cap.
> there is DC ground via the transformer winding.
That was a late revision. He may retain the original switching. Even if not, the transformer is maybe 1K DCR which leaves a couple dozen mV of DC, enough to pop.
The series cap ensures quite low DC on the output. The >2V 5u cap is so inexpensive it is worth keeping. A different thing from half-uFd at >400V.
We got 300V supply and only need a few Volts at the output jack. The proposed 0.47uFd is 10K at 35Hz. The both in series, 10K @ 40Hz. So the bottom note of bass needs ~~4V plate swing for 2V output. Not serious.
I started that way because the NFB loop leaks DC voltage to the output. This is rude. Hence the second cap.
> there is DC ground via the transformer winding.
That was a late revision. He may retain the original switching. Even if not, the transformer is maybe 1K DCR which leaves a couple dozen mV of DC, enough to pop.
The series cap ensures quite low DC on the output. The >2V 5u cap is so inexpensive it is worth keeping. A different thing from half-uFd at >400V.
If a 0µ47 is big enough for this application, why not.You get more distortion as the frequency drops, perhaps wanted ? But you can still leave out the second capacitor, with a transformer fixed in the circuit i see no pop.Only if you disconnect and connect again when the voltage is allready present.
Mona
+1 on the Ampeg suggestion. Note that there are some small differerences between the B15, B18, etc, that you might want to look into.
I also like the suggestion to use parallel 12AX7 preamp stages. That gives some added fatness to the tone.
I can't agree with the advice not to use a CF, as they do add some nice even harmonica warmth. BTW, if you have two CFs with an inverting gain stage between them, the output will limit symmetrically, rather than asymmetrically
Sent from my phone with Tapatalk. Please excuse any typpos.
I also like the suggestion to use parallel 12AX7 preamp stages. That gives some added fatness to the tone.
I can't agree with the advice not to use a CF, as they do add some nice even harmonica warmth. BTW, if you have two CFs with an inverting gain stage between them, the output will limit symmetrically, rather than asymmetrically
In this kind of application you probably want much higher 2nd harmonic. Research shows the human ear likes up to 20‰, perceiving it as warmth, rather than distortion.
Sent from my phone with Tapatalk. Please excuse any typpos.
Last edited:
> more distortion as the frequency drops
What is "more"? I worried it may be "too clean", in which case may as well use a chip.
For the post-tone stages I'm getting 2.8V peak with 0.1%THD @ 1K and 0.15%THD @ 50Hz, which meshes nicely with the 3dB-4dB cap loss down there. This is hardly a limit: 2% at 14Vpeak which is far beyond what a "preamp" should have to do. Note there will be 0.2%-4% THD in the first pre-tone stage, depending on level and gain.
However the plan shown has a high 6Hz bump which might need more study.
What is "more"? I worried it may be "too clean", in which case may as well use a chip.
For the post-tone stages I'm getting 2.8V peak with 0.1%THD @ 1K and 0.15%THD @ 50Hz, which meshes nicely with the 3dB-4dB cap loss down there. This is hardly a limit: 2% at 14Vpeak which is far beyond what a "preamp" should have to do. Note there will be 0.2%-4% THD in the first pre-tone stage, depending on level and gain.
However the plan shown has a high 6Hz bump which might need more study.
Some distorsion is probably not an issue here
About the "pop", i realised (a little late
) it is when using the output on the high side of the transformer.Well a capacitor in the feedback line (68k) will cure that (~1µ).I think the bump is a resonance from output capacitor and transformer inductance.
Mona
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.